Automated Progressive Delivery Using GitOps and Service Mesh

Intuit is using GitOps and Service Mesh to automate progressive delivery and improve anomaly detection in their infrastructure.

• Intuit faced challenges with seasonal service performance and traffic patterns when implementing progressive delivery.
• They are using Argo Rollouts in combination with Envoy and Service Mesh to provide automatic and declarative progressive delivery.
• Intuit is working on improving anomaly detection and observability using AI/ML-driven analysis.
• They are also exploring header-based routing for more intelligent canaries.
• Intuit's use of Service Mesh is driven by their zero-trust infrastructure and need for advanced traffic management.

Intuit uses header-based routing to do dark releases during tax season, allowing only Henrik's taxes to go through the new version.

Intuit is rolling out progressive delivery together with service mesh, using metrics from the mesh to automate analysis on the health of a new release. This is our next step towards more sophisticated techniques like AI/ML-driven analysis. As a FinTech Technology company with tens of millions of users and thousands of services, with many being very seasonal, we were facing a wide variety of service performance and traffic patterns that made patterning analysis templates challenging. This session discusses and demos how Intuit leverages Argo Rollouts in combination with Envoy and Service Mesh to provide automatic and declarative progressive delivery. Attendees will learn about the challenges we faced, our learnings and work we had to overcome those challenges Intuit has been heavily involved in building tools for progressive analysis with Argo Rollouts and we are now rolling this out at large scale together with service mesh. This is the story on how and what worked in production.Click here to view captioning/translation in the MeetingPlay platform!